# Research Scientist, Computer Vision · Good AI Labs

> Train visual models that give robots useful representations of objects, motion and physical interaction.

URL: https://www.goodailabs.com/careers/research-scientist-computer-vision/
Date: 2026-10-13
Last updated: 2026-10-13
Work arrangement: Remote-first


Related work: [Research at Good AI Labs](https://www.goodailabs.com/research/)

## Applying

Email your application to research@goodailabs.com, with the role title in the subject line. Include your location and availability. Please share only material you have permission to disclose.

Email: research@goodailabs.com. Include the role title in the subject line.


We are hiring a Research Scientist to develop visual learning for embodied intelligence. You will investigate how foundational models can learn representations of objects, motion and interaction that remain useful as a robot moves through a changing environment.

This role combines model research with practical training experience. The questions include what visual pretraining should capture, how temporal information should be represented and which learning signals produce useful transfer to physical tasks. You will shape a research direction and work with robotics researchers to evaluate the capabilities it enables.

## Responsibilities

- Develop and train visual models using images, video and observations from embodied tasks.
- Investigate representation learning and pretraining objectives for objects, motion, geometry and interaction.
- Study temporal information, partial visibility and changing viewpoints, with evaluations that test generalisation to unfamiliar scenes.
- Design data-selection and training experiments that isolate the causes of improved visual understanding.
- Connect visual evaluation to downstream policy performance, including the consequences of uncertainty and perceptual errors.
- Collaborate on efficient training and inference so promising visual capabilities can be used within a robot's resource constraints.

## Qualifications

- BS, MS or PhD in computer vision, machine learning, computer science or a related field, or demonstrated work of equivalent technical depth.
- Original work in computer vision, supported by hands-on experience training visual models and evaluating generalisation.
- A strong research record through first-author or leading work at venues such as CVPR, ICCV, ECCV, NeurIPS or ICLR, or substantial industrial or independent research of comparable depth.
- Practical knowledge of visual-data pipelines, augmentation, training objectives and optimisation, including debugging model collapse, overfitting or poor transfer.
- Strong Python and PyTorch skills, and sound foundations in geometry, linear algebra and statistical learning.
- The ability to identify a consequential visual-learning problem and carry it through a convincing implementation and evaluation.

## Preferred experience {#experience-that-would-help}

- Multimodal foundation models, self-supervised learning, video representation learning or embodied perception.
- Simulation and synthetic data for training visual models and testing robustness.
- Vision for manipulation, visual feedback in policy learning or inference under limited compute.

## Application materials {#what-to-include-in-your-application}

Include your CV and selected research contributions. For one visual model you trained, explain the data, learning objective, implementation work and evaluation. Papers, code, visual demonstrations and substantial proprietary projects can all establish the quality of your work.

Contact: research@goodailabs.com